This journal is© the Owner Societies 2017 Phys. Chem. Chem. Phys., 2017, 19, 5699--5704 | 5699

Cite this:Phys.Chem.Chem.Phys.,

2017, 19, 5699

Alpha-tocopherol inhibits pore formation inoxidized bilayers†

Phansiri Boonnoy,ab Mikko Karttunencd and Jirasak Wong-ekkabut*abe

In biological membranes, alpha-tocopherols (a-toc; vitamin E) protect

polyunsaturated lipids from free radicals. Although the interactions of

a-toc with non-oxidized lipid bilayers have been studied, their effects on

oxidized bilayers remain unknown. In this study, atomistic molecular

dynamics (MD) simulations of oxidized lipid bilayers were performed with

varying concentrations of a-toc. Bilayers with 1-palmitoyl-2-lauroyl-sn-

glycero-3-phosphocholine (PLPC) lipids and their aldehyde derivatives at

a 1 : 1 ratio were studied. Our simulations show that oxidized lipids self-

assemble into aggregates with a water pore rapidly developing across

the bilayer. The free energy of transporting an a-toc molecule in a bilayer

suggests that a-tocs can passively adsorb into it. When a-toc molecules

were present at low concentrations in bilayers containing oxidized lipids,

water pore formation was slowed down. At high a-toc concentrations,

no pores were observed. Based on the simulations, we propose that the

mechanism of how a-toc inhibits pore formation in bilayers with oxidized

lipids is the following: a-tocs trap the polar groups of the oxidized lipids

at the membrane–water interface resulting in a decreased probability of

the oxidized lipids making contact with the two leaflets and initiating

pore formation. This demonstrates that a-toc molecules not only protect

the bilayer from oxidation but also help to stabilize the bilayer after lipid

peroxidation occurs. These results will help in designing more efficient

molecules to protect membranes from oxidative stress.

Biological membranes serve as a partition between cells and theirenvironment. Under oxidative stress, unsaturated lipids present incell membranes may become exposed to attacks by free radicals,

that is, oxidation. Oxidation transforms some of the membranelipids to oxidized ones such as hydroperoxide and aldehyde lipids.1,2

It has also been suggested that internal, that is intra-leaflet, oxidationmay be important in altering bilayer properties.3

Lipid peroxidation is an important mechanism of cell membranedamage.4–6 Previous experiments and computer simulations4,7–14

have demonstrated how oxidized lipids disturb and deform bilayers.The retention of the polar chains of oxidized lipids in the bilayer’sinterior is energetically unfavorable resulting in the reversal of thepolar lipid chain to the bilayer interface.4,11,15,16 This reversal causesmajor changes in bilayer properties such as an increase in area perlipid, bilayer thinning, decrease of the lipid tail order parameter, andincrease in water permeability.4,12,15–18 Previously, we performed MDsimulations of lipid bilayers with oxidized lipids at high concentra-tions. Two major oxidized lipid species, hydroperoxide and aldehyde,were studied. The results showed that only aldehyde lipids were ableto induce pore formation across a PLPC bilayer and cause significantdeformation.13,15

a-toc is well-known as an efficient antioxidant that protectsmembranes from free radical-initiated oxidation.19–21 Naturalmembranes consist of saturated and unsaturated lipids, and they arepermeable to water and small molecules. Unsaturated lipids play animportant role in membrane permeability by disrupting the packingof saturated lipids. However, unsaturated lipids are readily suscep-tible to peroxidation which, if extreme, may lead to uncontrollabletransport of molecules across the membrane.22 Numerous studieshave been conducted to find the mechanisms of how a-toc protectspolyunsaturated fatty acids from free radicals and to explain howa-toc interacts with biological membranes.21,23–28 Protectivemechanisms in the absence of oxidized lipids have beenproposed, for example that the chromanol group of the a-tocmolecule can bind and trap free radicals within the interior andnear the membrane interface26 thus blocking their ability toenter the membrane and oxidize polyunsaturated lipid chains.

In previous studies, interactions of a-toc have been consideredwith non-oxidized lipid bilayers only24–28 and the effects of a-toc onbilayers with oxidized lipids remain unresolved. To understand howa-tocs interact with biological membranes after lipid peroxidation,

a Department of Physics, Faculty of Science, Kasetsart University, Bangkok 10900,

Thailand. E-mail: jirasak.w@ku.ac.thb Computational Biomodelling Laboratory for Agricultural Science and Technology

(CBLAST), Faculty of Science, Kasetsart University, Bangkok 10900, Thailandc Department of Mathematics and Computer Science & Institute for Complex

Molecular Systems, Eindhoven University of Technology, MetaForum,

5600 MB Eindhoven, The Netherlandsd Departments of Chemistry and Applied Mathematics, Western University,

1151 Richmond Street, London, Ontario N6A 5B7, Canadae Thailand Center of Excellence in Physics (ThEP Center),

Commission on Higher Education, Bangkok 10400, Thailand

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c6cp08051k

Received 24th November 2016,Accepted 24th January 2017

DOI: 10.1039/c6cp08051k

rsc.li/pccp

PCCP

COMMUNICATION

5700 | Phys. Chem. Chem. Phys., 2017, 19, 5699--5704 This journal is© the Owner Societies 2017

MD simulations of a-toc in oxidized lipid bilayers were carried out.Previous studies13,15 without a-toc have shown that the presence ofaldehyde lipids could lead to large disturbances of the bilayer.Water defects have been experimentally observed at 12.5–20 mol%of oxidation in 1-palmitoyl-2-oleoyl-sn-glycero-3-phosphocholine(POPC) membranes.18,29 In a computational study,17 water poreformation in the 1,2-dioleoyl-sn-glycero-3-phosphocholine (DOPC)bilayer was observed at high oxidation (75–100 mol%). Recently,our simulations13 with significantly longer simulation times(microseconds) suggested that the pore formation in oxidizedPLPC bilayers was unlikely to occur at aldehyde concentrationsbelow 50%. To achieve pore formation within reasonable computa-tional time, 1 : 1 binary lipid bilayer mixtures between PLPC and itstwo aldehyde lipids 1-palmitoyl-2-(9-oxo-nonanoyl)-sn-glycero-3-phosphocholine (9-al) and 1-stearoyl-2-(12-oxo-cis-9-dodecenoyl)-sn-glycero-3-phosphocholine (12-al) were chosen for this study.The molecular structures are shown in Fig. S1 (ESI†).

Methods

The simulated systems consisted of 0–16 a-toc moleculesembedded in lipid bilayers with 128 phospholipid moleculesand 10 628 simple point charge (SPC)30 water molecules. All thesimulated systems are listed in Table 1. Repeated simulationsfor each system were performed to verify reproducibility of theresults. The topologies and force field parameters of a-toc weretaken from Qin et al.24,25 The lipid parameters were taken fromprevious studies.4,13,31 Initially, the a-toc molecules were randomlyplaced in the water phase about 4.2 nm away from the center ofmass (COM) of the bilayer. After energy minimization, MD simula-tions were run for 1–2 ms with a 2 fs integration time step by usingthe GROMACS 4.5.5 package.32,33 Periodic boundary conditionswere applied in all directions. The neighbor list was updated atevery time step. A cutoff was employed at 1.0 nm for the real spacepart of electrostatic interactions and Lennard-Jones interactions.The Ewald particle-mesh34–36 was used to calculate the long-range part of electrostatic interactions and all bond lengthswere constrained by the LINCS algorithm.37 In all simulations,the temperature was set to 298 K using the v-rescale algorithm38

with a time constant of 0.1 ps. Pressure was controlled by the

Parrinello–Rahman algorithm39,40 with an equilibrium semi-isotropic pressure of 1 bar, a time constant of 4.0 ps andcompressibility of 4.5 � 10�5 bar�1. These parameters andprotocols have been extensively tested and optimized.41–43 Allvisualizations were carried out using Visual Molecular Dynamic(VMD) software.44

Free energy calculations

The potential of mean force (PMF) of an a-toc transferring intothe lipid bilayer was calculated using the umbrella samplingtechnique45 with the Weighted Histogram Analysis Method46

(WHAM). Three bilayers of 100% PLPC, 50% 12-al, and 50% 9-alwere used. A series of 41 simulations was run with the distancebetween a-toc and the bilayer center restrained between 0 and4.0 nm, with 0.1 nm spacing. In the first window, the a-tocmolecule is in bulk water. It was then subsequently moved intothe bilayer along the bilayer normal (z-axis) in each successivewindow. Therefore, the final window had the a-toc molecule atthe bilayer center. The hydroxyl of a-toc was restrained withrespect to the center of mass of the bilayer, using a harmonicrestraint with a force constant of 3000 kJ (mol nm2)�1 normal tothe bilayer. Simulations were performed in the NPT ensemble at298 K for a total time of 2.05 ms. Each window was run at least for50 ns and the last 20 ns was used for analysis. The bootstrapanalysis method47 was used to estimate the statistical uncertaintyin umbrella sampling simulations.

Results

The MD simulations show that without a-toc, oxidized lipidsself-assemble to form aggregates and a water pore developsrapidly across the bilayer (Fig. 1). A pore spanning the bilayeroccurred afterward. This result is in agreement with previousstudies.13 At low a-toc concentrations (2 and 4 a-toc moleculesin a bilayer), the a-toc molecules’ preferred position was close tothe bilayer interface which lead to slowed down formation (ascompared with no a-toc present) of a water pore over severalhundreds of nanoseconds, Table 1. Interestingly, when the a-toc

Table 1 Compositions of a-toc in the different oxidized lipid bilayers used in this study

No. System name Proportion Final structure Pore time (ns)

1 100% PLPC 128 PLPC Bilayer —2 100% PLPC + 1 a-toc 128 PLPC : 1 a-toc Bilayer —

3 50% 12-al 64 PLPC : 64 12-al Bilayer with a pore 140, 2384 50% 12-al + 2 a-toc 64 PLPC : 64 12-al : 2 a-toc Bilayer with a pore 337, 11005 50% 12-al + 4 a-toc 64 PLPC : 64 12-al : 4 a-toc Bilayer with a pore 340, 2666 50% 12-al + 8 a-toc 64 PLPC : 64 12-al : 8 a-toc Bilayer —7 50% 12-al + 16 a-toc 64 PLPC : 64 12-al : 16 a-toc Bilayer —

8 50% 9-al 64 PLPC : 64 9-al Bilayer with a pore 180, 509 50% 9-al + 2 a-toc 64 PLPC : 64 9-al : 2 a-toc Bilayer with a pore 212, 10010 50% 9-al + 4 a-toc 64 PLPC : 64 9-al : 4 a-toc Bilayer with a pore 809, 62011 50% 9-al + 8 a-toc 64 PLPC : 64 9-al : 8 a-toc Bilayer —12 50% 9-al + 16 a-toc 64 PLPC : 64 9-al : 16 a-toc Bilayer —

Communication PCCP

This journal is© the Owner Societies 2017 Phys. Chem. Chem. Phys., 2017, 19, 5699--5704 | 5701

concentration was increased, no pore formation was observedover the entire simulation time of over 2 ms.

Fig. 2 shows (unbiased MD simulation) that a-toc is able topassively penetrate into the bilayer and remain in the bilayerinterior around the carbonyl groups of the lipids. In agreementwith free energy calculations (Fig. 3), it is energetically favorable fora-toc to enter the bilayer. The free energy calculation shows that theadsorption energy of a-toc into a PLPC bilayer is �70.0 kJ mol�1

with the equilibrium position being about 1.3 nm from the centerof the bilayer. For moving the a-toc from equilibrium toward thecenter of the bilayer, an energy barrier with a steep slope andmagnitude of 13.6 kJ mol�1 was observed (Fig. 3). After reachingthe maximum of the free energy barrier, the PMF plateaus at

�56.8 kJ mol�1. Our PMF profile is qualitatively similar to that ofa cholesterol molecule transferring from water into a DPPC lipidbilayer.48 As a comparison, the free energies of cholesterol transfer-ring from water in DPPC lipid bilayers to the center of the bilayerand equilibrium are �50 and �67 kJ mol�1, respectively.48 Thisresult suggests that it is more favorable for an a-toc molecule to stayin lipid bilayer than cholesterol. Note that some of the quantitativedifferences might come from other parameters such as the type oflipid and temperature, and thus further studies are needed toestablish the differences more accurately.48

The free energy barrier of an a-toc between equilibrium andthe bilayer’s center was 13.6 kJ mol�1 (Fig. 3) and as a result,flip-flop of an a-toc in the pure PLPC bilayer was observed onlyonce. The PMF profiles of a-toc in oxidized bilayers are qualitativelysimilar to a non-oxidized bilayer but the adsorption energies wereincreased to �67.0 and �63.2 kJ mol�1 in 50% 12-al and 50% 9-albilayers, respectively. The equilibrium positions of the a-tocs weredeep inside the bilayer consistent with a decrease in the thicknessof the bilayer. These results suggest that a-toc molecules prefer tointeract with non-oxidized lipids and stay inside the non-oxidizedbilayer as compared to the oxidized system. Moreover, the freeenergy barriers from equilibrium toward the center of the bilayer of50% 12-al and 50% 9-al were 8.4 and 7.1 kJ mol�1, respectively. Theobserved decrease of the free energy barriers in oxidized bilayersresults in frequent flip-flops of a-toc (Fig. 4). Moreover, a-tocsalways formed hydrogen bonds with the aldehyde groups of theoxidized lipids’ tails. Flip-flops of a-tocs occurred when suchhydrogen bonds were lost and re-formed with aldehyde groupsin the opposite leaflet as shown in Fig. 4 and Fig. S2 (ESI†). Inmembranes containing no oxidized lipids, a-toc flip-flops havepreviously been observed only at the temperature of 350 K but

Fig. 1 Time evolution of a 50% 9-al lipid bilayer (A) without a-toc, (B) with4 a-toc molecules, and (C) with 8 a-toc molecules. Water molecules arenot shown for clarity. White, green, purple: PLPC, 9-al and a-toc mole-cules, respectively. Red: oxygen atoms of the a-toc hydroxyl groups. Thered circle represents the oxidized lipids’ aggregation region.

Fig. 2 The time evolution of the position of the hydroxyl group of ana-toc molecule in the 100% PLPC bilayer.

Fig. 3 The potential of mean force (PMF) of an a-toc transferring into thebilayer as a function of distance in the z-direction from the center of thebilayer (z = 0.0 nm). The curves show bilayers of 100% PLPC, 50% 12-al and50% 9-al in black, blue, and red lines, respectively. Bulk water defines zerofree energy. The dashed lines represent the average position of P atoms ineach bilayer. The free energies of a-toc adsorbing in the 100% PLPC, 50%12-al, and 50% 9-al bilayer are �70.0, �67.0, and �63.2 kJ mol�1,respectively. The free energy barriers of a-toc in the 100% PLPC, 50%12-al, and 50% 9-al bilayers are 13.6, 8.4, and 7.1 kJ mol�1, respectively.The maximum statistical uncertainties of 100% PLPC, 50% 12-al, and 50%9-al are 0.8, 0.4, and 0.3 kJ mol�1, respectively.

PCCP Communication

5702 | Phys. Chem. Chem. Phys., 2017, 19, 5699--5704 This journal is© the Owner Societies 2017

not below.24 These results suggest that a-toc is highly mobileinside bilayers containing oxidized lipids.

Electron density distributions (Fig. 5) show that the hydroxylgroups of a-toc molecules have their maxima at 0.96 nm and1.06 nm for the 50% 12-al and 50% 9-al systems, respectively.These maxima are related to the locations of the carbonylgroups of the lipid bilayers and are consistent with Fig. 4 andprevious studies.24,26–28 Surprisingly, the electron density of theoxygen atoms in oxidized lipids’ tails decreased at the bilayer’scenter and increased at the water interface when a-toc moleculeswere present. Our previous study suggested that one of the keymechanisms for passive pore formation is the distribution ofpolar groups inside the bilayer.13 The deep penetration of thepolar group inside the bilayer can bring lipids into contact withlipids in the opposite leaflet thus leading to the formation of awater bridge and consequently a stable pore.

When a-tocs are present in a bilayer, they are able to trap thepolar groups of the oxidized lipids at the water interface (Fig. 4and 5) resulting in a decreased probability of the oxidized lipidsbeing in contact with each other. Therefore, pores cannot beformed at high a-toc concentrations. Furthermore, water transport

across lipid bilayers with non-oxidized lipids is not frequent, sincethe free energy barrier of a water molecule crossing a PLPC bilayeris 29.4 � 2.3 kJ mol�1.4 Permeability also increases by an order ofmagnitude as the concentration of oxidized lipids increases.17,18

Moreover, permeability increases by three orders of magnitudewhen a pore spans the bilayer (Table 2, Fig. S3, ESI†). When theconcentration of a-toc inside the bilayer increases, water perme-ability decreases and pore formation becomes inhibited (Table 2).

The deuterium order parameter for the sn-1 chain of PLPCand oxidized lipids (Fig. S4, ESI†) shows that all systems of lipidbilayers are in the biologically relevant liquid crystalline phaseand the order parameter increases with increasing a-toc concen-tration. Cholesterol and a-toc molecules have similar molecularstructures consisting of a hydrophobic tail and a ring-structurewith a hydroxyl group. The effects of these molecules on bio-logical membranes are similar and this has been suggested to beresponsible for the observed increase in bilayer thickness andlipid tail order.25,49,50 Moreover, Issack et al.51 have shown that

Fig. 4 (A) The time evolution of the position of the hydroxyl group ofa-toc from the center of the 50% 9-al lipid bilayer (along the z-axis). (B) Thenumber of hydrogen bonds between the hydroxyl groups of a-tocs andthe aldehyde groups of the oxidized lipid tails in the upper and lowerleaflets. This result is from only one of eight a-toc molecules in the systemof 50% 9-al with 8 a-toc and the rest of the a-toc molecules are shown inFig. S2 (ESI†).

Fig. 5 Electron density profiles for the oxygen atoms in the oxidized lipidhydrocarbon chains with and without a-toc, and for the hydroxyl groups ofa-toc in 50% 12-al and 50% 9-al bilayers. The bilayers with a-toc consistedof 16 a-toc molecules.

Table 2 Water permeability through the oxidized lipid bilayer

No. System name

Bilayer PorePermeability(pf) cm3 s�1 (�10�15)

Permeability(pf) cm3 s�1 (�10�12)

1 100% PLPC 0.96 � 0.24 —2 100% PLPC + 1 a-toc 0.44 � 0.16 —

3 50% 12-al 6.00 � 1.94 7.31 � 0.314 50% 12-al + 2 a-toc 3.27 � 0.59 4.72 � 0.595 50% 12-al + 4 a-toc 3.61 � 0.85 5.89 � 0.406 50% 12-al + 8 a-toc 4.22 � 0.53 —7 50% 12-al + 16 a-toc 2.81 � 0.45 —

8 50% 9-al 4.50 � 1.65 8.97 � 0.349 50% 9-al + 2 a-toc 3.72 � 0.87 8.28 � 0.4310 50% 9-al + 4 a-toc 3.55 � 0.53 6.06 � 1.4711 50% 9-al + 8 a-toc 2.21 � 0.48 —12 50% 9-al + 16 a-toc 2.57 � 0.45 —

Note: Water permeability was calculated using pf = nwRt/NA, where nw isthe average volume of a single water molecule, 18 cm3 mol�1, Rt is therate of water transport across the bilayer and NA is Avogadro’s number.

Communication PCCP

This journal is© the Owner Societies 2017 Phys. Chem. Chem. Phys., 2017, 19, 5699--5704 | 5703

the free-energy barrier for transferring a water molecule to thecenter of the bilayer was increased by 6 kJ mol�1 when 41 mol%cholesterol was present in a DPPC bilayer. This observed significantincrease in the free energy barrier to transfer water through thebilayer with the presence of cholesterol results in a decrease inwater permeability.51,52

In conclusion, free radicals play an important role in membranedamage and aldehyde lipids are the major oxidative lipid productthat causes pore formation and bilayer deformation.13,19–21 On theother hand, a-toc is one of the most effective antioxidants inremoving free radicals and it has been used in cosmetics, func-tional foods and many other applications.53–55 Previously,19–21,26

the only protective mechanism of a-toc against lipid peroxidation inbiological membranes was proposed to be due to a-toc blockingfree radicals’ entry into the membrane thus protecting polyun-saturated lipid chains from oxidation processes, that is, with nooxidized lipids in the bilayer. In a realistic case, however, oxidizedlipids are present and it is important to understand how theirdestructive effects can be prevented. In this study, we have shownthat a-toc molecules can inhibit pore formation in oxidized lipidbilayers by confining the polar groups of the oxidized lipids at thewater interface. Our findings also suggest that by controllinga-toc concentration, the stability of biological membranes canbe increased. This understanding of how a-tocs affect oxidizedmembranes is likely to be beneficial for designing new mole-cules to protect, e.g., skin, against aging56,57 and in plasmatreatment of cancer.49

Competing financial interests

The authors declare no competing financial interest.

Acknowledgements

Financial support by Kasetsart University Research & Develop-ment Institute (KURDI; J. W.) and Faculty of Science (J. W.) isacknowledged.

References

1 T. K. Mandal and S. N. Chatterjee, Radiat. Res., 1980, 83,290–302.

2 S. N. Chatterjee and S. Agarwal, Free Radical Biol. Med.,1988, 4, 51–72.

3 P. L. Else and E. Kraffe, Biochim. Biophys. Acta, 2015, 1848,417–421.

4 J. Wong-ekkabut, Z. Xu, W. Triampo, I. M. Tang, D. PeterTieleman and L. Monticelli, Biophys. J., 2007, 93, 4225–4236.

5 H. Y. Yin, L. B. Xu and N. A. Porter, Chem. Rev., 2011, 111,5944–5972.

6 E. Niki, Y. Yoshida, Y. Saito and N. Noguchi, Biochem.Biophys. Res. Commun., 2005, 338, 668–676.

7 X. M. Li, R. G. Salomon, J. Qin and S. L. Hazen, Biochemistry,2007, 46, 5009–5017.

8 J. P. Mattila, K. Sabatini and P. K. Kinnunen, Langmuir,2008, 24, 4157–4160.

9 K. Sabatini, J. P. Mattila, F. M. Megli and P. K. Kinnunen,Biophys. J., 2006, 90, 4488–4499.

10 L. Beranova, L. Cwiklik, P. Jurkiewicz, M. Hof and P. Jungwirth,Langmuir, 2010, 26, 6140–6144.

11 H. Khandelia and O. G. Mouritsen, Biophys. J., 2009, 96,2734–2743.

12 V. Jarerattanachat, M. Karttunen and J. Wong-ekkabut,J. Phys. Chem. B, 2013, 117, 8490–8501.

13 P. Boonnoy, V. Jarerattanachat, M. Karttunen and J. Wong-ekkabut, J. Phys. Chem. Lett., 2015, 6, 4884–4888.

14 P. Siani, R. M. de Souza, L. G. Dias, R. Itri and H. Khandelia,Biochim. Biophys. Acta, 2016, 1858, 2498–2511.

15 L. Cwiklik and P. Jungwirth, Chem. Phys. Lett., 2010, 486,99–103.

16 J. Garrec, A. Monari, X. Assfed, L. M. Mir and M. Tarek,J. Phys. Chem. Lett., 2014, 5, 1653–1658.

17 M. Lis, A. Wizert, M. Przybylo, M. Langner, J. Swiatek,P. Jungwirth and L. Cwiklik, Phys. Chem. Chem. Phys.,2011, 13, 17555–17563.

18 K. A. Runas and N. Malmstadt, Soft Matter, 2015, 11,499–505.

19 E. Niki and M. G. Traber, Ann. Nutr. Metab., 2012, 61,207–212.

20 X. Wang and P. J. Quinn, Prog. Lipid Res., 1999, 38, 309–336.21 M. G. Traber and J. Atkinson, Free Radical Biol. Med., 2007,

43, 4–15.22 M. H. Brodnitz, J. Agric. Food Chem., 1968, 16, 994–999.23 V. E. Kagan, Ann. N. Y. Acad. Sci., 1989, 570, 121–135.24 S. S. Qin, Z. W. Yu and Y. X. Yu, J. Phys. Chem. B, 2009, 113,

16537–16546.25 S. S. Qin and Z. W. Yu, Acta Phys.-Chim. Sin., 2011, 27,

213–227.26 X. L. Leng, J. J. Kinnun, D. Marquardt, M. Ghefli, N. Kucerka,

J. Katsaras, J. Atkinson, T. A. Harroun, S. E. Feller andS. R. Wassall, Biophys. J., 2015, 109, 1608–1618.

27 D. Marquardt, N. Kucerka, J. Katsaras and T. A. Harroun,Langmuir, 2014, 31, 4464–4472.

28 D. Marquardt, J. A. Williams, N. Kucerka, J. Atkinson,S. R. Wassall, J. Katsaras and T. A. Harroun, J. Am. Chem.Soc., 2013, 135, 7523–7533.

29 R. Volinsky, L. Cwiklik, P. Jurkiewicz, M. Hof, P. Jungwirthand P. K. Kinnunen, Biophys. J., 2011, 101, 1376–1384.

30 H. J. C. Berendsen, J. P. M. Postma, W. F. van Gunsteren andJ. Hermans, in Intermol. Forces, ed. B. Pullman, D. Reidel,Dordrecht, The Netherlands, 1981, pp. 331–342.

31 M. Bachar, P. Brunelle, D. P. Tieleman and A. Rauk, J. Phys.Chem. B, 2004, 108, 7170–7179.

32 B. Hess, C. Kutzner, D. van der Spoel and E. Lindahl,J. Chem. Theory Comput., 2008, 4, 435–447.

33 S. Pronk, S. Pall, R. Schulz, P. Larsson, P. Bjelkmar, R. Apostolov,M. R. Shirts, J. C. Smith, P. M. Kasson, D. van der Spoel, B. Hessand E. Lindahl, Bioinformatics, 2013, 29, 845.

34 T. Darden, D. York and L. Pedersen, J. Chem. Phys., 1993, 98,10089–10092.

35 U. Essmann, L. Perera, M. L. Berkowitz, T. Darden, H. Leeand L. G. Pedersen, J. Chem. Phys., 1995, 103, 8577–8593.

PCCP Communication

5704 | Phys. Chem. Chem. Phys., 2017, 19, 5699--5704 This journal is© the Owner Societies 2017

36 M. Karttunen, J. Rottler, I. Vattulainen and C. Sagui, Curr.Top. Membr., 2008, 60, 49.

37 B. Hess, H. Bekker, H. J. C. Berendsen and J. G. E. M. Fraaije,J. Comput. Chem., 1997, 18, 1463–1472.

38 G. Bussi, D. Donadio and M. Parrinello, J. Chem. Phys., 2007,126, 014101.

39 M. Parrinello and A. Rahman, J. Appl. Phys., 1981, 52,7182–7190.

40 S. Nose and M. L. Klein, Mol. Phys., 1983, 50, 1055–1076.41 J. Wong-Ekkabut, M. S. Miettinen, C. Dias and M. Karttunen,

Nat. Nanotechnol., 2010, 5, 555–557.42 J. Wong-ekkabut and M. Karttunen, J. Chem. Theory Com-

put., 2012, 8, 2905–2911.43 J. Wong-ekkabut and M. Karttunen, Biochim. Biophys. Acta,

2016, 1858, 2529.44 W. Humphrey, A. Dalke and K. Schulten, J. Mol. Graphics,

1996, 14, 33–38.45 G. M. Torrie and J. P. Valleau, J. Comput. Phys., 1977, 23,

187–199.46 S. Kumar, J. M. Rosenberg, D. Bouzida, R. H. Swendsen,

P. A. Kollman and J. M. Rosenbergl, J. Comput. Chem., 1992,13, 1011–1021.

47 J. S. Hub, F. K. Winkler, M. Merrick and B. L. de Groot,J. Am. Chem. Soc., 2010, 132, 13251–13263.

48 W. F. D. Bennett, J. L. MacCallum, M. J. Hinner, S. J.Marrink and D. P. Tieleman, J. Am. Chem. Soc., 2009, 131,12714–12720.

49 J. Van der Paal, E. C. Neyts, C. C. W. Verlackt andA. Bogaerts, Chem. Sci., 2016, 7, 489–498.

50 C. Hofsass, E. Lindahl and O. Edholm, Biophys. J., 2003, 84,2192–2206.

51 B. B. Issack and G. H. Peslherbe, J. Phys. Chem. B, 2015, 119,9391–9400.

52 H. Saito and W. Shinoda, J. Phys. Chem. B, 2011, 115,15241–15250.

53 J. M. Tucker and D. M. Townsend, Biomed. Pharmacother.,2005, 59, 380–387.

54 A. Azzi, R. Gysin, P. Kempna, R. Ricciarelli, L. Villacorta,T. Visarius and J. M. Zingg, Mol. Aspects Med., 2003, 24,325–336.

55 Institute of Medicine, Vitamin E, Dietary Reference Intakesfor Vitamin C, Vitamin E, Selenium, and Carotenoids, TheNational Academies Press, Washington, DC, 2000.

56 F. H. Lin, J. Y. Lin, R. D. Gupta, J. A. Tournas, J. A. Burch,M. A. Selim, N. A. Monteiro-Riviere, J. M. Grichnik,J. Zielinski and S. R. Pinnell, J. Invest. Dermatol., 2005,125, 826–832.

57 J. Krutmann, Hautarzt, 2011, 62, 576.

Communication PCCP